www.nature.com/scientificreports

OPEN Bioinformatics analysis of DNMT1 expression and its role in head and neck squamous cell carcinoma prognosis Jili Cui1,2,4, Lian Zheng2,3,4, Yuanyuan Zhang1,2* & Miaomiao Xue1,2*

Head and neck squamous cell carcinoma (HNSCC) is the sixth most common type of malignancy in the world. DNA cytosine-5-methyltransferase 1 (DNMT1) play key roles in carcinogenesis and regulation of the immune micro-environment, but the gene expression and the role of DNMT1 in HNSCC is unknown. In this study, we utilized online tools and databases for pan-cancer and HNSCC analysis of DNMT1 expression and its association with clinical cancer characteristics. We also identifed genes that positively and negatively correlated with DNMT1 expression and identifed eight hub genes based on protein–protein interaction (PPI) network analysis. Enrichment analyses were performed to explore the biological functions related with of DNMT1. The Tumor Immune Estimation Resource (TIMER) database was performed to explore the relationship between DNMT1 expression and immune-cell infltration. We demonstrated that DNMT1 gene expression was upregulated in HNSCC and associated with poor prognosis. Based on analysis of the eight hub genes, we determined that DNMT1 may be involved in cell cycle, proliferation and metabolic related pathways. We also found that signifcant diference of B cells infltration based on TP 53 mutation. These fndings suggest that DNMT1 related epigenetic alterations have close relationship with HNSCC progression, and DNMT1 could be a novel diagnostic biomarker and a promising therapeutic target for HNSCC.

Head and neck squamous cell carcinoma (HNSCC), an aggressive malignancy that is difcult to diagnose early, accounts for more than 95% of head and neck ­cancers1,2. Most HNSCC patients are diagnosed at an advanced stage with an extremely poor prognosis. It is estimated that 400,000 new patients are diagnosed annu- ally worldwide­ 3. Surgical resection, along with radiotherapy and chemotherapy, is recommended for early stage HNSCC. Although a comprehensive and multidisciplinary approach, including chemotherapy, radiotherapy, and immunotherapy, are used for advanced-stage HNSCC, the 5-year overall survival (OS) rate has remained below 50%4,5. Tus, by delineating the mechanisms of tumorigenesis and immune infltration in HNSCC, a foundation for future therapeutic developments will be laid. 5-methylcytosine (m5C) modifcation is an important post-transcriptional methylation, and m5C is associ- ated with the carcinogenesis and tumorigenesis of various ­cancers6. 5-methylcytosine regulators function as one kinds of important epigenetic ­modifers6 . DNA methylation plays a role in epigenetic gene silencing, X chromosome inactivation, and genome stability­ 7. It has also been shown that DNA methylation modifcations are crucial for normal cell cycle regulation, development, and diferentiation­ 8. Recent studies have demonstrated that epigenetic DNA methylation alterations drive formation of many types of ­cancer9. DNA methylation is catalyzed by DNA methyltransferase enzymes (DNMTs), which regulate epigenetic modifcations in mammals. DNMTs methylate CpG islands in promoters, aberrant modifcation of DNMTs methylate could cause carcinogenesis and tumor progression­ 10. DNA cytosine-5-methyltransferase 1 (DNMT1) acts as a crucial epigenetic modifer by regulating DNA methylation of target genes, which involves in tumor suppressor gene ­expression11. DNMT1 gene expression is also subject to epigenetic regulation. Numerous studies have shown that DNMTI is highly upregulated in various cancers such as cholangiocarcinoma­ 12, hematological malignancies­ 13, pancreatic cancer­ 14, breast ­cancer15. DNMTI is required for maintaining the cancer stem cell ­status16,17. In lung cancer patients,

1Department of General Dentistry, The First Afliated Hospital, Zhengzhou University, No. 1 Jianshe Road, Zhengzhou 450052, Henan, China. 2Key Laboratory of Clinical Medicine, The First Afliated Hospital, Zhengzhou University, Zhengzhou 450052, Henan, China. 3Department of Oral and Maxillofacial Surgery, The First Afliated Hospital, Zhengzhou University, Zhengzhou 450052, Henan, China. 4These authors contributed equally: Jili Cui and Lian Zheng. *email: [email protected]; [email protected]

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 1 Vol.:(0123456789) www.nature.com/scientificreports/

Figure 1. High DNMT1 expression was observed in head and neck cancers and in pan-cancers (A) High DNMT1 expression was found in several types of cancers. (B) Te DNMT1 transcript level was signifcantly higher in HNSCC tissue than in non-tumor tissue. (C) DNMT1 protein expression was higher in HNSCC tumor tissue than in non-tumor tissue.

high levels of DNMT1 expression are associated with poor clinical prognosis­ 18.However, it is also reported that upregulation of DNMT1 also functions in tumor-suppressor gene activation­ 19,20. Further, DNMT1 regulates the immune system and is indispensable for the inhibitory function of Foxp3 + Treg cells­ 21. In our previous study, we utilized the bioinformation analysis and validated the characters of m5C related regulators in HNSCC, and DNMT1 functioned as a key regulator in HNSCC progression­ 22. We identifed that DNMT1 might serve as a crucial m5C regulator in malignant activities of HNSCC, and increased DNMT1 expression was associated with patient’s mortality ­rate22. Nevertheless, the comprehensively biological activities of DNMT1 in HNSCC tumorigenesis and progression has not yet been clarifed. In this study, we used public databases and online analysis websites to unravel the role of DNMT1 in HNSCC tumorigenesis and progression. Furthermore we identifed methylation markers with diagnostic potential, and identifed potential targets for the treatment of HNSCC. Results DNMT1 gene expression and mutation distribution. We performed a pan-cancer analysis of DNMT1 expression, and we found that DNMT1 was upregulated in several diferent tumor types (Fig. 1A). Specifcally, transcription level of DNMT1 was signifcantly higher in HNSCC tumor tissues (n = 520) than in normal tissues (n = 44, p < 0.001) (Fig. 1B). To further validate the expression of DNMT1, the Immunohistochemistry (IHC) was applied to detect the protein expression of DNMT1 gene in normal tissues and tumor tissues based on HPA. Results showed that DNMT1 protein expression was signifcantly overexpressed in HNSCC tissues compared with normal tissue (Fig. 1C). Tus, DNMT1 had potential as a biomarker for diagnosis of HNSCC.

The clinical features and prognosis of DNMT1 in HNSCC. To explore the DNMT1 related clinical characters, we determined whether gender, age, race, tumor stage, tumor grade, nodal metastasis, and human papillomavirus (HPV) infection status correlated with DNMT1 expression in HNSCC. We found that DNMT1 expression level was signifcantly higher in males patients than in females patients (p < 0.001) (Fig. 2A) and in the less than 60-year-old subgroups (n = 256), DNMT1 expression level was remarkably higher compared with over61-year-old subgroups (n = 261, p = 0.003) (Fig. 2B). We also observed that DNMT1 expression was higher in Caucasians compared with Asians and African-American subgroups (p = 0.037) (Fig. 2C). Advanced-grade tumors had higher DNMT1 gene expression level compared with lower grade tumor samples (Fig. 2D), but

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 2 Vol:.(1234567890) www.nature.com/scientificreports/

Figure 2. Relationships between DNMT1 expression and clinical characteristics and prognosis were analyzed. (A–G) DNMT1 expression correlated with gender, age, race, tumor grade, and HPV infectious status, but not with lymphatic invasion or tumor stage. (H, I) OS and RFS analyses demonstrated that patients with high DNMT1 expression had poor clinical prognosis.

TNM stages and lymphatic invasion did not correlate with DNMT1 expression (Fig. 2E,F). We validated that HPV-positive patients (n = 41) had higher DNMT1 expression compared with HPV negative patients (n = 80) (p < 0.001) (Fig. 2G). Furthermore, we explored the overall survival (OS) and relapse-free survival (RFS) of HNSCC patients (Fig. 2H,I) and found that patients with high DNMT1 expression had better OS rate (HR = 0.69, logrank p = 0.0076). Patients with high expression level of DNMT1 showed no signifcant diference in RFS com- pared with low DNMT1 expression subgroups (HR = 2.4, logrank p = 0.094). DNMT1 may be potential genes for HNSCC poor prognosis prediction and may provide reference value for HNSCC treatment in the further.

DNMT1 gene mutations and co‑expressed genes. Te somatic mutation rate of DNMT1 in HNSCC patients was 1.6%, and the eight mutations identifed were all missense mutations, P1330S, P1325S, E912Q, S1352G, P692S, H370Y, T616M, and R325L (Fig. 3A). We also identifed eight co-related genes such as CLSPN, UHRF1, BRCA1, ATAD5, TIMELESS, CIT, KIF4B and DTL (Table 1). Te correlation between DNMT1 mutation status and the 8 genes mutation status were evaluated. In this study, the 8 genes showed signifcantly strong correlation with DNMT1 expression, as illustrated in Fig. 3B. CLSPN (Spearman: 0.78, p = 3.46e−102), UHRF1 (Spearman: 0.75, p = 2.11e−89), BRCA1(Spearman: 0.74, p = 6.34e−86), ATAD5 (Spearman: 0.73, p = 4.92e−84), TIMELESS (Spearman: 0.73, p = 3.52e−82), CIT (Spear- man: 0.71, p = 2.51e−76), KIF4B (Spearman: 0.71, p = 2.93e−75), and DTL (Spearman: 0.70, p = 4.90e−74).

DNMT1 expression associates with immune cell infltration and HNSCC prognosis. Tumor- infltrating immune cells are an independent predictor of cancer therapy and prognosis. We investigated the correlation between DNMT1 expression and infltration of various immune cells in HNSCC, and we found that DNMT1 expression weakly correlated with immune cell purity (Correlation = 0.199, p = 8.49e−06). And the

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 3 Vol.:(0123456789) www.nature.com/scientificreports/

Figure 3. Distribution of DNMT1 mutations and co-expressed genes. (A) In HNSCC, the somatic mutation rate was 1.6%. (B) CLSPN, UHRF1, BRCA1, ATAD5, TIMELESS, CIT, KIF4B, and DTL genes were co-expressed with DNMT1.

Correlated Gene Cytoband Spearman’s correlation p-value CLSPN 1p34.3 0.783 3.46E−102 UHRF1 19p13.3 0.75 2.11E−89 BRCA1 17q21.31 0.741 6.34E−86 ATAD5 17q11.2 0.735 4.92E−84 TIMELESS 12q13.3 0.73 3.25E−82 CIT 12q24.23 0.711 2.51E−76 KIF4B 5q33.2 0.707 2.93E−75 DTL 1q32.3 0.703 4.90E−74

Table 1. Te co-expression genes of DNMT1 in HNSCC.

top three immune cells that correlated with DNMT1 expression in HNSCC were CD4 + T cells (Cor = 0.412), neutrophils (Cor = 0.328), and dendritic cells (Cor = 0.360) (Fig. 4A). To further investigate associations with immune-cell infltration status, HNSCC patients were divided into HPV-negative and HPV-positive subgroups. Te results illustrated that HPV-negative subgroup had mildly correlation with immune purity and immune cells infltration such as CD4 + T cells (Cor = 0.429, p = 4.00e−19), macrophage (Cor = 0.25, p = 5.17e−07), neu- trophil (Cor = 0.301, p = 1.17e−09), dendritic cell (Cor = 0.331, p = 1.46e−11) (Fig. 4B,C). Furthermore, we iden- tifed the immune cells infltration with TP53 mutation status, the results demonstrated that the TP53 mutation had signifcant diference in B cells, CD8 + Tcells, neutrophil, dendritic cells infltration, and weak diference in CD4 + T cells infltration (Fig. 4D). We also validated the correlation between the immune-cell infltration status and prognosis in HNSCC. In this study, we demonstrated that the cumulative survival in B cell high expres- sion group showed weak diference compared with low expression group (Log-rank p = 0.045) (Fig. S1a). In HPV-positive group, immune cells high expression groups had better cumulative survival compared with low expression patients, such as in B cell (Log-rank p = 0.013), CD8 + T cell (Log-rank p = 0.015), neutrophil (Log- rank p = 0.074) (Fig. S1b). However, in HPV-negative group the cumulative survival in diferent immune cells infltration showed no signifcantly diference (Fig. S1c). In this part, our fnding indicated that DNMT1 expres- sion have closely relationship with immune cells infltration.

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 4 Vol:.(1234567890) www.nature.com/scientificreports/

Figure 4. Te relationship between DNMT1 expression and immune cell infltration and clinical prognosis of HNSCC. (A) DNMT1 gene expression positively correlated with tumor purity and infltration of B cells, CD8 + T cells, CD4 + T cells, macrophages, neutrophils, and dendritic cells in HNSCC. (B) DNMT1 expression did not correlate with tumor purity and macrophage infltration in the HPV-positive HNSCC group. Infltration of B cells, CD8 + T cells, CD4 + T cells, neutrophils, and dendritic cells correlated weakly with DNMT1 expression in the HPV-positive HNSCC subgroup. (C) DNMT1 expression positively correlated with tumor purity and infltration of CD8 + T cells, CD4 + T cells, macrophages, neutrophils, and dendritic cells, but not with infltration of B cells, in the HPV-negative HNSCC subgroup. (D) TP53 mutation with immune cells infltration in HNSC.

Identifcation and analysis of DEGs. To explore the DNMT1 related diferently expressed genes, we identifed 100 positively correlated and 100 negatively correlated genes as illustrated in Fig. 5A. To clarify the top correlated genes, we mapped the top 8 genes such as CLSPN, TIMELESS, ATAD5, UHRF1, BRCA1, BRIP1, TMPO, and CENPF, which were positively correlated with DNMT1. POLD4, GUK1, TMED3, TOM1, SEPW1, SERF2, TMEN208, VAMP8, MYEVO2, and ATP5EP2 were top 10 signifcantly negative correlated genes with DNMT1 expression (Fig. 5B), and the map showed that DNMT1 expression associated with multiple genes expression.

Enrichment analysis of DNMT1‑related genes. To identify the pathways that the DNMT1-related genes are involved in, we performed GO and KEGG analyses. As shown in Fig. 6A, Te GO BP enrichment results showed that the DNMT1-related genes associate with DNA replication, transcription initiation, and microtubule-based movement. GO CC analysis showed a signifcant association with nucleoplasm activity (Fig. 6B), and GO MF analysis showed a strong association with ATP binding (Fig. 6C). Furthermore, KEGG results showed remark associations with the cell cycle and DNA replication (Fig. 6D). Overall, DNMT1 appears to function in basic cellular activities such as, DNA replication initiation, nucleoplasm, ATP binding, and cell cycle.

PPI network and hub gene selection of DNMT1. To further investigate the function of DNMT1 in HNSCC, we constructed a PPI network based on the top 100 positively correlated DNMT1-related genes, and we found that the top 100 positive DNMT1-related genes correlated with each other (Fig. 7A). We identifed eight hub genes such as BUB1B, KIF11, CDC20, MCM2, MCM4, TOP2A, ASPM and MAD2L1 (Fig. 7B) and we detected the correlations between hub genes and DNMT1 expression (Fig. 7C). Moreover, we found that expres- sion levels of all eight hub genes were signifcantly increased in HNSCC tissue compared with normal tissue (all p < 0.001) (Fig. S2a). To determine the overall survival among these 8 genes, we applied Kaplan–Meier analysis,

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 5 Vol.:(0123456789) www.nature.com/scientificreports/

Figure 5. Distribution of genes that associated with DNMT1 expression. (A) Te volcano map shows the genes that positively and negatively correlated with DNMT1 expression. (B) Te heat-map shows the top 10 genes that positively correlated and negatively with DNMT1 expression.

and results showed that high expression level of ASPM in HNSC had better prognosis (HR = 0.71 (0.51–0.98), logrank p = 0.035). Te low expression level of BUB1B in HNSCC was associated with poor prognosis (HR = 1.38 (0.94–1.78), logrank p = 0.046) (Fig. S2b). Tus, the hub genes have potential as novel targets for delineation of the tumor progression mechanism in HNSCC. Discussion HNSCC has a high mortality rate and is difcult to detect at an early stage due to the lack of efective early moni- toring and screening ­factors2. Once HNSCC is diagnosed, it is already advanced and the prognosis is extremely poor. It has been found that traditional surgery, postoperative radiotherapy, and chemotherapy have not signif- cantly improve survival of advanced ­HNSCC1. Te recent emergence of immunotherapy may ultimately improve the prognosis of advanced-stage HNSCC patients, but substantial research is still needed. Te epigenetic alterations contain the reversible DNA methylation and histone acetylation or ­methylation23. DNA methylation modifcation is an important regulator of cell activity, and abnormal regulation due to altera- tions in methylation can lead to tumorigenesis­ 15. DNMTs are involved in initiation of mammalian chromatin remodeling and gene expression regulation in diverse physiological and pathophysiological ­processes23,24. m5C functions as an important epigenetic modifer in DNA and RNA with transcriptional expression­ 25. m5C plays vital roles in HNSCC carcinogenesis and tumor progression­ 22. In addition, genomic alterations in m5C-related modifers have close relationship with clinicopathological characteristics. In our previous study, we found that the DNMT1, a key m5C regulator, was closely related with HNSCC prognosis, and also identifed the DNMT1 as death risk predictor in HNSCC progression. DNMT1 is an important regulator of methylation, since it participates in the regulation of various key biologi- cal activities and the immune cell micro-environment26,27. However, the biological roles and potential mecha- nisms of DNMT1 expression remain largely unknown in HNSCC. Insights into the mechanism of DNMT1 expression will provide a foundation to determine the role of DNMT1 in HNSCC and may lead to improvements in the clinical management and diagnosis of HNSCC patients. It has been shown that dysregulation of DNMT1 associates with tumorigenesis and progression of multiple types of cancer. Overexpression of DNMT1 has been observed in macroadenomas, invasive tumors, and advanced pituitary ­adenoma28. DNMT1 has been shown to be involved in drug resistance of non-small cell cancer cells, and this drug resistance was reversed upon treatment with a micro-RNA that negatively regulates DNMT1 expression­ 29,30. Similarly, increased DNMT1 expression has been found in cancerous mammary gland tissues of a mammary cancer animal model. DNMT1 is also expressed highly in difuse large B-cell lymphoma­ 31. Based on previous bioinformation analysis of HNSCC and m5C regulators, we comprehensively explored the m5C regulator DNMT1 and its roles in HNSCC carcinogenesis and progression process. Consistent with previous studies, we also found that DNMT1 is expressed highly in HNSCC tissue compared with normal tissue. Pan- cancer analysis also showed upregulation of DNMT1 transcription in several cancer types. We found that high DNMT1 expression associated with HNSCC patients ≥ 60 years old, which is consistent with a previous study in which high DNMT1 expression was observed in peripheral blood mononuclear cells (PBMCs) from patients up to 64 years old in a large European cohort­ 32. Previous studies have shown that genetic alterations in DNMT1 associated with a low risk of developing breast cancer in the European Caucasian popula- tion and that DNMT1 expression was higher in Caucasians than in Asians and African-Americans33. We also demonstrated that HPV-positive HNSCC was more malignant and associated with poor prognosis. DNMT1 gene expression was also positively correlated with poor prognosis, suggesting that DNMT1 expression could be a potential biomarker of HNSCC prognosis. Te somatic mutation rate of DNMT1 was 1.6%, and all of the mutations were missense mutations. Our results demonstrated that DNMT1 mutations may cause genomic instability leading to tumorigenesis.

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 6 Vol:.(1234567890) www.nature.com/scientificreports/

Figure 6. GO enrichment and KEGG pathways analyses of the top 100 genes that positively correlated with DNMT1 expression. (A) GO BP enrichment analysis of DNMT1-associated genes. (B) GO CC enrichment analysis of DNMT1-associated genes. (C) Go MF enrichment analysis of DNMT1-associated genes. (D) KEGG analysis of DNMT1-associated genes.

TP53 is a suppressor gene with high frequency of mutation in tumors, and has become an independent prog- nostic factor in tumorigenesis. TP53 mutation can signifcantly activate immune checkpoints, initiate efector T cells and increase immune factors expression levels­ 34. In lung cancer immunotherapy, TP53 mutation can be recognized as a predictor of immunotherapy sensitivity­ 35. However, TP53 mutational status based immune responses and the potential clinical characters involves in HNSCC has been evaluated. Terefore, it is necessary to further investigate the meaningful immune-related prognostic models and immune-related efects of TP53 status to provide powerful prognostic biomarkers and increase the efectiveness of ­immunotherapy36. Recent studies have shown that immune-cell infltration contributes to the tumor immune micro-environment, which afects ­immunotherapy37,38. In this study, we demonstrated that HNSCC patients with abundant infltration of immune cells, such as B cells, CD8 + T cells, neutrophils, had better prognosis. Previous studies have shown that high levels of infltration by CD8 + tumor-infltrating lymphocytes and B cells led to better clinical outcomes and prognosis. Terefore, the immune-cell landscape of HNSCC may associate with HNSCC clinical parameters. In our study, DNMT1 was upregulated in various cancers, especially in HNSCC tumor tissues compared with adjacent tissues. And DNMT1 expression was much higher in male patients, and higher in elder patients over 60-year old. Te advanced stages of HNSCC had higher DNMT1 transcription levels compared with early stages. Furthermore, we also investigated the DNMT1 related co-expression genes, such as CLSPN, UHRF1, BRCA1, ATAD5, TIMELESS, CIT, KIF4B and DTL. In addition, the associated top 3 immune cells infltrations of DNMT1 were as follows, CD4 + T cells (Cor = 0.412), neutrophils (Cor = 0.328), and dendritic cells (Cor = 0.360). We also

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 7 Vol.:(0123456789) www.nature.com/scientificreports/

Figure 7. Te PPI network was based on the top 100 DNMT1-related genes. (A) Te distribution of the top 100 DNMT1-related genes is shown, and the red-colored nodes represent the log (FC) values of the Z-scores of the proteins that interact with the designated protein. (B) 8 hub genes were identifed with Cytoscape sofware. Te red-color represents genes that positively correlated with DNMT1 expression. (c) Correlations between expression of hub genes and DNMT1 expression.

comprehensively explored the distribution of DNMT1 related DGEs, pathways enrichment analysis and the PPI network in HNSCC. Terefore, we identifed DNMT1 as a potential biomarker for HNSCC. Further studies are needed to uncover the underlying mechanism of DNMT1 in tumorigenesis to develop biomarkers for HNSCC early detection and prognosis. Material and methods Clinical datasets and DNMT1 expression. Te Cancer Genome Atlas (TCGA) database provide com- prehensive molecular characterizations of multiple cancers and comprehensive data on gene profling and sequencing, as ­described39–42. Tese bioinformatic computational tools were used to identify novel therapeutic and diagnostic biomarkers among 520 HNSCC tumor tissues and 44 adjacent tissues. Te gene transcription profles and clinical information f or these tissue samples were included in this study. Samples from patients with incomplete information were excluded. Te Ualcan database (http://ualca​n.path.uab.edu/analy​sis.html) is a web-portal platform to further explore TCGA gene expression data­ 43. Pan-cancer analysis of DNMT1 expres- sion and its relationship with general clinical information were performed via Ualcan. We also used Human Protein Atlas (HPA, http://www.prote​inatl​as.org/), a free and open source website used to investigate protein alterations and expression. DNMT1 mutation and co-expression information was obtained from the cbioportal database (https​:// www.cbiop​ortal​.org/), and the Linkomics database (http://linke​domic​s.org/) was used to identify diferentially expressed genes (DEGs) of ­DNMT144,45.

Tumor immune estimation resource (TIMER) database. Te TIMER ­database46 (Version 1.0 1, https​://cistr​ome.shiny​apps.io/timer​/) is a public database used to explore the relationship between cancer and

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 8 Vol:.(1234567890) www.nature.com/scientificreports/

immune-cell infltration. TP53 mutation appears to be the most frequent genetic alteration overall. In order to clarify the immune cells infltration with TP53 mutation in HNSCC, we utilized TIMER to evaluate the impact of immune factors and DNMT1 on HNSCC prognosis and to evaluate the efect of TP53 gene expression on immune cell infltration.

Survival of HNSCC patients. Te Kaplan–Meier plotter online ­tool47 (https​://kmplo​t.com/analy​sis/) was used to assess the correlation between DNMT1 gene expression and HNSCC survival. Statistical signifcance was determined by hazard ratio (HR) with a 95% confdence interval (CI) and log-rank p-value.

Protein–protein interaction (PPI) network analysis and hub gene selection. LinkedOmics (http://linke​domic​s.org/) was also performed to identify correlated genes­ 44. A gene activity network was con- structed with the top 100 genes that positively correlated with DNMT1 expression. Te top eight genes that positively correlated with DNMT1 expression were used in enrichment analyses for hub genes and the pro- tein–protein interaction (PPI) network. Te PPI network of overlapped DEGs was constructed by STRING and visualized with Cytoscape. Te method has been described in previous study­ 48.

Gene function enrichment analyses. GO (Gene Ontology) and KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway enrichment analyses were used to evaluate function. GO analysis assesses biological processes (BP), molecular functions (MF), and cellular components (CC). KEGG pathway enrichment analysis was performed using the online website DAVID as described in previous study­ 49.

Statistical analysis. Te model construction and statically methods have been described as previous ­studies50–52. Te absolute values 0.00–0.19, 0.20–0.39, 0.40–0.59, 0.60–0.79, and 0.80–1.0 indicated very weak, weak, moderate, strong, and very strong correlations, respectively. p < 0.05 was considered statistically signif- cant.

Received: 25 October 2020; Accepted: 14 January 2021

References 1. Cramer, J., Burtness, B., Le, Q. & Ferris, R. Te changing therapeutic landscape of head and neck cancer. Nat. Rev. Clin. Oncol. 16, 669–683. https​://doi.org/10.1038/s4157​1-019-0227-z (2019). 2. Karam, S. & Raben, D. Radioimmunotherapy for the treatment of head and neck cancer. Lancet Oncol. 20, e404–e416. https://doi.​ org/10.1016/s1470​-2045(19)30306​-7 (2019). 3. Shirai, K. & O’Brien, P. Molecular targets in squamous cell carcinoma of the head and neck. Curr. Treat. Options Oncol. 8, 239–251. https​://doi.org/10.1007/s1186​4-007-0030-4 (2007). 4. Kaidar-Person, O., Gil, Z. & Billan, S. Precision medicine in head and neck cancer. Drug Resist. Update 40, 13–16. https​://doi. org/10.1016/j.drup.2018.09.001 (2018). 5. Oliva, M. et al. Immune biomarkers of response to immune-checkpoint inhibitors in head and neck squamous cell carcinoma. Ann. Oncol. 30, 57–67. https​://doi.org/10.1093/annon​c/mdy50​7 (2019). 6. He, Y., Zhang, Q., Zheng, Q., Yu, X. & Guo, W. Distinct 5-methylcytosine profles of circular RNA in human hepatocellular carci- noma. Am. J. Transl. Res. 12, 5719–5729 (2020). 7. Teschendorf, A. & Relton, C. Statistical and integrative system-level analysis of DNA methylation data. Nat. Rev. Genet. 19, 129–147. https​://doi.org/10.1038/nrg.2017.86 (2018). 8. Greenberg, M. & Bourc’his, D. Te diverse roles of DNA methylation in mammalian development and disease. Nat. Rev. Mol. Cell Biol. 20, 590–607. https​://doi.org/10.1038/s4158​0-019-0159-6 (2019). 9. Michalak, E., Burr, M., Bannister, A. & Dawson, M. Te roles of DNA, RNA and histone methylation in ageing and cancer. Nat. Rev. Mol. Cell Biol. 20, 573–589. https​://doi.org/10.1038/s4158​0-019-0143-1 (2019). 10. DNMT3A DNA-binding residues provide specifcity for CpG DNA methylation. Cancer Discov. 8, 14, https://doi.org/10.1158/2159-​ 8290.Cd-rw201​8-028 (2018). 11. Chen, Z. & Zhang, Y. Role of mammalian DNA methyltransferases in development. Annu. Rev. Biochem. 89, 135–158. https://doi.​ org/10.1146/annur​ev-bioch​em-10301​9-10281​5 (2020). 12. Colyn, L. et al. Dual targeting of G9a and DNMT1 for the treatment of experimental cholangiocarcinoma. Hepatology (Baltimore, Md.), https​://doi.org/10.1002/hep.31642​ (2020). 13. Hoang, N. & Rui, L. DNA methyltransferases in hematological malignancies. J. Genet. Genomics (Yi chuan xue bao) 47, 361–372. https​://doi.org/10.1016/j.jgg.2020.04.006 (2020). 14. Wong, K. DNMT1 as a therapeutic target in pancreatic cancer: Mechanisms and clinical implications. Cell. Oncol. (Dordr.) 43, 779–792. https​://doi.org/10.1007/s1340​2-020-00526​-4 (2020). 15. Wong, K. DNMT1: A key drug target in triple-negative breast cancer. Semin. Cancer Biol. https​://doi.org/10.1016/j.semca​ ncer.2020.05.010 (2020). 16. 16Ferreri, A., Illerhaus, G., Zucca, E. & Cavalli, F. Flows and faws in primary central nervous system lymphoma. Nat. Rev. Clin. Oncol. 7, https​://doi.org/10.1038/nrcli​nonc.2010.1039-c1031​; author reply https​://doi.org/10.1038/nrcli​nonc.2010.1039-c1032​, https​://doi.org/10.1038/nrcli​nonc.2010.9-c1 (2010). 17. Zagorac, S. et al. DNMT1 inhibition reprograms pancreatic cancer stem cells via upregulation of the miR-17-92 cluster. Cancer Res. 76, 4546–4558. https​://doi.org/10.1158/0008-5472.Can-15-3268 (2016). 18. Piperi, C., Vlastos, F., Farmaki, E., Martinet, N. & Papavassiliou, A. Epigenetic efects of lung cancer predisposing factors impact on clinical diagnosis and prognosis. J. Cell. Mol. Med. 12, 1495–1501. https​://doi.org/10.1111/j.1582-4934.2008.00309​.x (2008). 19. Lin, R. et al. Te tobacco-specifc carcinogen NNK induces DNA methyltransferase 1 accumulation and tumor suppressor gene hypermethylation in mice and lung cancer patients. J. Clin. Investig. 120, 521–532. https​://doi.org/10.1172/jci40​706 (2010). 20. Kim, J. et al. Aberrant methylation of the FHIT gene in chronic smokers with early stage squamous cell carcinoma of the lung. Carcinogenesis 25, 2165–2171. https​://doi.org/10.1093/carci​n/bgh21​7 (2004).

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 9 Vol.:(0123456789) www.nature.com/scientificreports/

21. Wang, L. et al. Foxp3+ T-regulatory cells require DNA methyltransferase 1 expression to prevent development of lethal autoim- munity. Blood 121, 3631–3639. https​://doi.org/10.1182/blood​-2012-08-45176​5 (2013). 22. Xue, M. et al. Gene signatures of m5C regulators may predict prognoses of patients with head and neck squamous cell carcinoma. Am. J. Transl. Res. 12, 6841–6852 (2020). 23. Turek-Plewa, J. & Jagodziński, P. Te role of mammalian DNA methyltransferases in the regulation of gene expression. Cell. Mol. Biol. Lett. 10, 631–647 (2005). 24. Xue, C., Zhao, Y. & Li, L. Advances in RNA cytosine-5 methylation: Detection, regulatory mechanisms, biological functions and links to cancer. Biomark. Res. 8, 43. https​://doi.org/10.1186/s4036​4-020-00225​-0 (2020). 25. Chellamuthu, A. & Gray, S. Te RNA methyltransferase NSUN2 and its potential roles in cancer. Cells https​://doi.org/10.3390/ cells​90817​58 (2020). 26. Lal, G. & Bromberg, J. Epigenetic mechanisms of regulation of Foxp3 expression. Blood 114, 3727–3735. https://doi.org/10.1182/​ blood​-2009-05-21958​4 (2009). 27. Jeltsch, A. & Jurkowska, R. Allosteric control of mammalian DNA methyltransferases—A new regulatory paradigm. Nucleic Acids Res. 44, 8556–8575. https​://doi.org/10.1093/nar/gkw72​3 (2016). 28. Ma, H. et al. Overexpression of DNA (cytosine-5)-methyltransferase 1 (DNMT1) and DNA (cytosine-5)-methyltransferase 3A (DNMT3A) is associated with aggressive behavior and hypermethylation of tumor suppressor genes in human pituitary adenomas. Med. Sci. Monit. 24, 4841–4850. https​://doi.org/10.12659​/msm.91060​8 (2018). 29. Chhabra, R. miRNA and methylation: A multifaceted liaison. Chembiochem Eur. J. Chem. Biol. 16, 195–203. https://doi.org/10.1002/​ cbic.20140​2449 (2015). 30. Wang, P. et al. Kindlin-2 interacts with and stabilizes DNMT1 to promote breast cancer development. Int. J. Biochem. Cell Biol. 105, 41–51. https​://doi.org/10.1016/j.bioce​l.2018.09.022 (2018). 31. Loo, S., Ab Hamid, S., Musa, M. & Wong, K. DNMT1 is associated with cell cycle and DNA replication gene sets in difuse large B-cell lymphoma. Pathol. Res. Pract. 214, 134–143, https​://doi.org/10.1016/j.prp.2017.10.005 (2018). 32. Ciccarone, F. et al. Age-dependent expression of DNMT1 and DNMT3B in PBMCs from a large European population enrolled in the MARK-AGE study. Aging Cell 15, 755–765. https​://doi.org/10.1111/acel.12485​ (2016). 33. Kullmann, K., Deryal, M., Ong, M., Schmidt, W. & Mahlknecht, U. DNMT1 genetic polymorphisms afect breast cancer risk in the central European Caucasian population. Clin. Epigenet. 5, 7. https​://doi.org/10.1186/1868-7083-5-7 (2013). 34. Dong, Z. et al. TP53 potential predictive value of and mutation status for response to PD-1 blockade immunotherapy in lung adenocarcinoma. Clin. Cancer Res. 23, 3012–3024. https​://doi.org/10.1158/1078-0432.Ccr-16-2554 (2017). 35. Skoulidis, F. et al. STK11/LKB1 mutations and PD-1 inhibitor resistance in -mutant lung adenocarcinoma. Cancer Discov. 8, 822–835. https​://doi.org/10.1158/2159-8290.Cd-18-0099 (2018). 36. Ham, S. et al. TP53 gain-of-function mutation promotes infammation in glioblastoma. Cell Death Difer. 26, 409–425. https://doi.​ org/10.1038/s4141​8-018-0126-3 (2019). 37. Schabath, M. et al. Diferential association of STK11 and TP53 with KRAS mutation-associated gene expression, proliferation and immune surveillance in lung adenocarcinoma. Oncogene 35, 3209–3216. https​://doi.org/10.1038/onc.2015.375 (2016). 38. Liu, R., Hu, R., Zeng, Y., Zhang, W. & Zhou, H. Tumour immune cell infltration and survival afer platinum-based chemotherapy in high-grade serous ovarian cancer subtypes: A gene expression-based computational study. EBioMedicine 51, 102602. https​:// doi.org/10.1016/j.ebiom​.2019.10260​2 (2020). 39. Xue, C. et al. Low expression of LACTB promotes tumor progression and predicts poor prognosis in hepatocellular carcinoma. Am. J. Transl. Res. 10, 4152–4162 (2018). 40. He, Y. et al. Prediction of hepatocellular carcinoma prognosis based on expression of an immune-related gene set. Aging 12, 965–977. https​://doi.org/10.18632​/aging​.10266​9 (2020). 41. Ye, C. et al. Whole-genome DNA methylation and hydroxymethylation profling for HBV-related hepatocellular carcinoma. Int. J. Oncol. 49, 589–602. https​://doi.org/10.3892/ijo.2016.3535 (2016). 42. Rajendran, V., Gopalakrishnan, C. & Sethumadhavan, R. Pathological role of a point mutation (T315I) in BCR-ABL1 protein-A computational insight. J. Cell. Biochem. 119, 918–925. https​://doi.org/10.1002/jcb.26257​ (2018). 43. Chandrashekar, D. et al. UALCAN: A portal for facilitating tumor subgroup gene expression and survival analyses. Neoplasia 19, 649–658. https​://doi.org/10.1016/j.neo.2017.05.002 (2017). 44. Vasaikar, S., Straub, P., Wang, J. & Zhang, B. LinkedOmics: Analyzing multi-omics data within and across 32 cancer types. Nucleic Acids Res. 46, D956–D963. https​://doi.org/10.1093/nar/gkx10​90 (2018). 45. Zhang, G., Kang, Z., Mei, H., Huang, Z. & Li, H. Promising diagnostic and prognostic value of six genes in human hepatocellular carcinoma. Am. J. Transl. Res. 12, 1239–1254 (2020). 46. Li, T. et al. TIMER2.0 for analysis of tumor-infltrating immune cells. Nucleic Acids Res. 48, W509–W514, https://doi.org/10.1093/​ nar/gkaa4​07 (2020). 47. Menyhárt, O., Nagy, Á. & Győrfy, B. Determining consistent prognostic biomarkers of overall survival and vascular invasion in hepatocellular carcinoma. R. Soc. Open Sci. 5, 181006. https​://doi.org/10.1098/rsos.18100​6 (2018). 48. Doncheva, N., Morris, J., Gorodkin, J. & Jensen, L. Cytoscape StringApp: Network analysis and visualization of proteomics data. J. Proteome Res. 18, 623–632. https​://doi.org/10.1021/acs.jprot​eome.8b007​02 (2019). 49. Nie, H., Mei, J., Zhang, Q., An, F. & Zhan, Q. Systematic characterization of the expression and prognostic values of formin-like gene family in gastric cancer. DNA Cell Biol. https​://doi.org/10.1089/dna.2020.5508 (2020). 50. Bhardwaj, V. & Purohit, R. Structural changes induced by substitution of amino acid 129 in the coat protein of Cucumber mosaic virus. Genomics 112, 3729–3738. https​://doi.org/10.1016/j.ygeno​.2020.04.023 (2020). 51. Bhardwaj, V. & Purohit, R. A new insight into protein-protein interactions and the efect of conformational alterations in PCNA. Int. J. Biol. Macromol. 148, 999–1009. https​://doi.org/10.1016/j.ijbio​mac.2020.01.212 (2020). 52. Singh, R., Bhardwaj, V., Das, P. & Purohit, R. Natural analogues inhibiting selective cyclin-dependent kinase protein isoforms: A computational perspective. J. Biomol. Struct. Dyn. 38, 5126–5135. https​://doi.org/10.1080/07391​102.2019.16967​09 (2020). Acknowledgements Tis research was funded by the Science and Technology Research Project of Henan Province (202102310074). Author contributions J.C., L.Z.: design the study, analysis and interpretation of data. Y.Z. acquisition and analysis of the data; M.X. conception, design, drafing of the work. Interpretation of the data.

Competing interests Te authors declare no competing interests.

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 10 Vol:.(1234567890) www.nature.com/scientificreports/

Additional information Supplementary Information Te online version contains supplementary material available at https​://doi. org/10.1038/s4159​8-021-81971​-5. Correspondence and requests for materials should be addressed to Y.Z. or M.X. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Open Access Tis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. Te images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/.

© Te Author(s) 2021

Scientifc Reports | (2021) 11:2267 | https://doi.org/10.1038/s41598-021-81971-5 11 Vol.:(0123456789)